Human-in-the-Loop Optimization for Inclusive Design: Balancing Automation and Designer Expertise

Abstract

Accessible and inclusive design has gained increased attention in HCI, yet practical implementation remains challenging due to resource-intensive prototyping methods. Traditional approaches such as workshops, A-B tests, and co-design sessions struggle to capture the diverse and complex needs of users with disabilities at scale. This position paper argues for an automated, accessible Human-in-the-Loop (HITL) design optimization process that shifts the designer's role from directly crafting prototypes to curating constraints for algorithmic exploration. By pre-constraining the design space based on specific user interaction needs, integrating adaptive multi-modal feedback channels, and personalizing feedback prompts, the HITL approach could efficiently refine design parameters, such as text size, color contrast, layout, and interaction modalities, to achieve optimal accessibility. This approach promises scalable, individualized design solutions while raising critical questions about constraint curation, transparency, user agency, and ethical considerations, making it essential to discuss and refine these ideas collaboratively at the workshop.